Category Archives: BLOG

Appraising Transitional Justice Through the Just War Theory Analogue


Steven R. Ratner is Bruno Simma Collegiate Professor of Law at the University of Michigan. His teaching and research focus on public international law and on a range of challenges facing governments and international institutions since the Cold War, including territorial disputes, counter-terrorism strategies, ethnic conflict, state and corporate duties regarding foreign investment, and accountability for human rights violations.


Colleen Murphy’s The Conceptual Foundations of Transitional Justice offers a significant addition to our understanding of transitional justice (TJ) in at least three ways.

First, Murphy deserves accolades for traversing numerous disciplinary divides in her approach to TJ. For scholars seeking to build more bridges between international law and philosophy, this books offers an excellent way as to how it can be done. Although legal scholarship on TJ has recognized many of the goals that Murphy defends, lawyers will benefit from the careful way in which she discusses trade-offs and priorities. Philosophers, on the other hand, should learn the importance of direct engagement with the key legal sources, including the workings of several TJ mechanisms.

Second, Murphy’s explanation of the justice of TJ moves us beyond both political and legal ways of thinking about TJ. Students of politics tend to think about TJ in terms of its success or failure for the particular country, with different measurements of those terms. Lawyers take the “justice” in TJ to mean that the starting point must be criminal justice, although they realize that TJ must encompass other goals like truth, acknowledgment, reparations, and measures for non-recurrence. They also tend to zero in on the scope of a state’s duties under international law (e.g., are amnesties allowed? Is there a right to the truth?)

Murphy helps us see that TJ is a matter of justice as philosophers understand that concept. She also emphasizes the necessity of defining clearly the subject matter of justice – that the justice of one thing (e.g., the basic structure of a liberal state) is quite different from the justice of something else (e.g., a TJ course of action taken by a state). Chapter 1 thus offers up the criteria of a TJ situation, so we know why the justice of that thing is worthy of inquiry. The four criteria that she offers seem to encompass the real-world cases that we ordinarily think of as concerning TJ and to raise morally important questions.

Third, and the focus of my comment, Murphy adopts an original methodological hook for analyzing the justice of a state’s TJ choices, namely to examine separately the just goals of TJ and the just means of carrying it out. Her analogy is to the two branches of just war theory (and the of law of war as well), namely the jus ad bellum (JAB) and the jus in bello (JIB). She thus develops what she calls the jus ad bellum analogue (JABA) and the jus in bello analogue (JIBA). This bifurcation of justice has the potential to provide a new level of clarity in our thinking, as we accept the need to look at two aspects of the process and not merely one.

At the same time, the JABA/JIBA approach raises some new conceptual questions. First, within the JABA itself, although the goal of social transformation (ST) – which she equates with relational transformation (RT) – is morally compelling, is that goal the only — or the principally — just goal for TJ? To return to the legal paradigm, in the case of the JAB, we can say that the only legally defensible reasons for a state to use force against another state are self-defense, authorization from the Security Council, and, on some accounts, humanitarian intervention. It is a necessary and sufficient condition of JAB legality that the state action be based on one of these three grounds. But does ST/RT meet the necessary/sufficient test? Though Murphy shows that ST through RT can respond to the four unique moral problems of a TJ situation, it is not clear if ST is the best or only means to address those problems. And even if it is the best or only method, why does that make it just in the same sense that JAB tells us the necessary and sufficient conditions for a just (or legal) war?

Second, and relatedly, because many TJ situations arise after prolonged violence in a country (not just “normalized and collective wrongdoing”), one wonders what role in the JABA is played by solidification of domestic peace. The serious existential uncertainty that helps define a situation needing TJ can extend to whether the peace will even hold. Murphy points out (p.111) that stability has its own value, as people still need to live their lives, with school, businesses, marriages, and ordinary living continuing. Yet those seem to take a back seat to ST.

Certainly, elites in a transitional situation can use the canard of solidification of the peace as a way to avoid any serious ST. I’ve witnessed this myself when I served on two UN accountability panels, for Cambodia and Sri Lanka, where the triumphalist elites running those countries each insisted that trials, or even investigations, of international crimes would re-ignite civil wars (by the Khmer Rouge and the LTTE, respectively) that were clearly over. But certainly it seems morally compelling goal of TJ to ensure that peace is maintained in those situations where it might be precarious. ST via RT can provide a way to long-term domestic peace, but it is not at all clear that it can provide short-term peace. Transformation may have to be put on hold to placate the losers or those with the ability to upset the transition.

Third, the JABA/JIBA tool can give rise to some new questions regarding a just TJ. International law doctrine generally holds that JAB and JIB operate independently. Under the so-called “separation thesis,” the legality of going to war is not affected by how it is carried out, and vice-versa. In particular, a state that obeys the rules of war (JIB) cannot be relieved of its liability for an aggressive war (JAB); and states that use force for legally acceptable reasons (JAB) still must obey the rules of war (JIB), even against aggressors. This principle explains the practice of the International Committee of the Red Cross of not commenting on the underlying legality of a war, as it insists that all parties must comply with the JIB.

Yet JAB and JIB are not actually completely independent. After states banned most uses of force in Article 2(4) of the Charter, states adjusted the JIB in the 1949 Geneva Conventions to make clear that territory annexed by the occupier was still legally occupied. And as most states began to accept that anti-colonial wars were lawful, they changed the jus in bello in the 1977 Additional Protocol I to state that anti-colonial fighters who did not resemble traditional soldiers would also receive most of the privileges of combatantcy and not be treated as illegal fighters. (See Addition Protocol I, arts. 4, 44). So it turns out that the legality of the recourse to force can affect the legality of the conduct of the war.

In Murphy’s use of the JIBA/JABA model, the thrust of her Chapter 4 seems based on the separation thesis – that we judge the justice of the manner of a transition independently of its goals. She thus usefully says TJ responses must be “fitting or apt” (p. 163), which turns on four features (pp. 163-72). First, they must address six moral imperatives, i.e., they must respond to the perpetrators by (1) repudiating their wrongs and (2) holding them accountable; they must respond to the victims by (3) acknowledging the wrongs against them, (4) recognizing their status as victims, and (5) providing reparation; and they must respond to both by (6) promoting nonrecurrence. These correspond to goals widely accepted by practitioners of TJ in international organizations and NGOs.  Second, they must reflect the relations among the relevant actors. Third, they must reflect the gravity of the offense. And fourth, they must reflect cultural expectations. She goes on to show, through good examples, how, different institutional responses – e.g., the ICTY and the Ugandan Amnesty Commission – did not “fit” based on those criteria. And she astutely observes the importance for states of “active coordination” by domestic and international actors to make sure that each TJ mechanism is considered in connection with the others, rather than in isolation.

Yet I wonder whether, like the law, strict separation does not offer a complete description of the two aspects of justice. In particular, is there a connection between the JIBA requirements and the JABA requirement? It would be nice to explore how each of the imperatives regarding perpetrators and victims advances ST/RT or how the imperatives might change if the goals were different from ST/RT. I agree with her inclination, which she justifies in the conclusion, not to ask how “any specific kind of response” (p. 198) actually contributes to ST, but it would help to know how the six imperatives (which are one level up in generality) do so.

Second, in thinking about the six moral imperatives for TJ responses, one might ask why are they not also – or instead – part of the JABA, i.e., the just goals of a transition? International organizations like the UN and other scholars of TJ generally describe those six imperatives as part of the end, in part because ST/RT sounds too academic as well as idealistic, but also because they see those imperatives as part of the goals of TJ. So are these six imperatives intrinsically valuable, regardless of their actual effect on ST/RT? Intrinsically valuable but also instrumentally valuable?

Murphy ends with an important plea for a holistic view of justice, one sensitive to the JABA and all the many elements of the JIBA. I certainly agree, although maybe the international organizations have it right by having us consider justice only from the latter perspective. If a state finds the proper mechanisms that meet the six moral goals in a way that that also respects the other three JIBA criteria, does it matter whether its goal is ST/RT or something else? One could consider that a great success and also be confident that societal transformation would follow on its own.   At the same time, by justifying ST/RT as the ultimate long-term just solution to the unique problems of TJ, Murphy has helped us see where the state eventually needs to go.

Fragility, Authority and the Ethics of Transitions


Nir Eisikovits is an Associate Professor of Philosophy & Director of the Applied Ethics Center at the University of Massachusetts Boston.  His areas of expertise include Transitional Justice and Post War Reconstruction, Ethics of War, and International Relations Theory.


Colleen Murphy’s new book on transitional justice displays her signature blend of analytic rigor, elegant writing and empirically anchored theorizing. She follows up her excellent first book on political reconciliation with a volume on what it means to transform a war torn society so that it can, ultimately, become reconciled. The just pursuit of political transformation, Murphy argues, is at the heart of the idea of transitional justice. This is the best, most ambitious philosophical account of transitional justice that I have read. The book can be read with great utility by scholars and students seeking to understand the unique conditions and dilemmas surrounding transitions, as well as by policy makers interested in fashioning decent and legitimate transitional institutions.

In this essay I will focus on Murphy’s characterization of the unique circumstances holding during times of transitions. To develop a normative account of transitional justice one needs to understand the conditions that countries struggling to make a start after war or mass atrocity face. This is why Murphy’s book must begin with laying these conditions out.

Murphy tells us that transitional states are characterized by pervasive structural inequality – a legacy of the unjust social arrangements that held in the past. Consider, for example, how apartheid misshaped the life prospects of Blacks in South Africa and the resulting deep inequalities. These states also suffer from a normalization of collective and political wrongdoing (to stay with the South African example, consider that for many in that country the apartheid state – through its institutions and actors – was seen as an agent of harm). It made your husband disappear, it enlisted your sister to spy on her own people, it made you worry about the safety of your children.

Under such circumstances, Murphy reminds us, major crimes become normalized – an expected part of life’s fabric: “wrongdoing such as rape, disappearing or torture… becomes a basic fact of life for individuals in the midst of conflict … a fact around which individuals must orients their conduct” (55). The third feature of transitions is “serious existential uncertainty”: political instability, lack of resources and the lingering influence and power of the old guard make countries that try to emerge from prolonged violence especially fragile. We just don’t know if they will make it through and many of them don’t, or at least fail to make it through as democracies (consider Egypt’s post Arab Spring turmoil, or even the authoritarian version of Rwanda that emerged under Kagame). Finally, there is fundamental uncertainty about authority: does a transitional regime have the political, legal and moral authority to “rule and enforce rules?” (72) and does it have the authority to address past wrongs and work towards social transformation – especially if those wrongs were legal at the time of commission and if the new regime is not completely purged of those who were influential in the past?

I would like to further flesh out some of these conditions – especially the last two – and suggest some implications for political transitions. The fundamental uncertainty about authority and political fragility attendant to transitions are exacerbated by a lack of political traditions and the lack of a shared political history and identity. In settled democracies both leaders and citizens can appeal to a store of past experiences and some settled views or traditions about how to deal with extreme circumstances. These traditions can provide guidance and a background against which to judge current conduct, even when it is extreme or unprecedented (in fact, the very ability to agree that the conduct of an official is unprecedented and completely strays from widely accepted traditions can buttress existing institutions and put current turmoil in context).

A fledgling United States considering John Adams’ prolonged absences from the capital and his temper tantrums in a very different way from a modern United States assessing President Trump’s behavior. In the intervening two and a half centuries the country gradually developed a set of expectations and traditions about how its leader should act. An observer of Adams could legitimately ask whether that was what a president was supposed to do (ironically, as Vice President, Adams who was very aware of the fragility of the institutions of the new republic and tried to invest the presidency with grandeur and gravitas by proposing various titles by which the president should be known. His efforts were unappreciated, ultimately earning him the moniker “his rotundity”). An observer of President Trump’s antics has an answer to that question. Stated differently, it is easier for settled democracies to get through periods of significant political fragility. What makes these hardships less existentially frightening in settled polities is the existence of political traditions, shared history, some sense, if you will, of political identity, that can put threats in context (“we’ve been through worse…” “don’t panic – remember how many people supported Nixon in the first few months after his maleficence was made public …” “we have a self correcting political system” and so on).

Transitional polities, then, are precarious because, on the one hand, so much hangs on the success of their transitional processes and, on the other, they have little guidance and very few tools to successfully shepherd themselves through. Spain immediately after Franco was often steps away from falling back into a dictatorship. South Africa after the demise of apartheid was similarly close to the brink. And in each case these countries were pretty much flying blind – dependent on the political instincts of their leaders, international good will, the exhaustion of their citizens, and, frequently, dumb luck. Under these circumstances, one wonders how much we can expect, morally, from a process of political transformation. Murphy invokes a fascinating analogy to Just War Theory and its distinction between Jus ad Bellum (the justice of the decision to go to war) and Jus in Bello (the justice of the war’s conduct), to remind us that a transition is subject to two layers of moral judgment: first, whether it seeks to create the right institutions and instill the right principles and, second, on how it does these things. But the combination of political fragility and lack of clear lines of authority which Murphy describes, raise questions about whether transitions to democracy can really be pursued democratically and about the moral meaning of a failure to do so.

Spain’s transition was facilitated by a pacto del olvido or pact of forgetting – a refusal, for many decades, to talk about the horrors of the civil war and the crimes committed in its aftermath. South Africa’s transition was facilitated by the work of its Truth and Reconciliation Commission (TRC). Both of these, in very different ways, were failures to live up to standards of the rule of law. The Spanish Pacto for the obvious reason that it completely sidelined the need of victims for public acknowledgement; the TRC because its hearings jettisoned basic tenets of the rule of law and because, to paraphrase Michael Ignatieff, it inaugurated South African democracy by letting a bunch of murderers get away with murder. And yet there are reasons to give both of these states a moral pass. Not a permanent pass; it is, of course, much more problematic to insist on silence in Spain now than it was in the 1970’s, and South Africa cannot continue to build its legal system on truth for amnesty arrangements. But given the extreme fragility, high stakes and lack of legal and political standards to guide action – we tend to view these “sins” of transition leniently. Murphy, even though she considers the different ways transitional policies can fail the “jus in bello” test, does not quite consistently consider the possibility that governments engaged in transitions are structurally set up to fail it or to seriously risk failing it.

Compartmentalizing Transitional Justice


In August this year, three French investigating judges were appointed to assess allegations that the Bank BNP Paribas was complicit in the Rwandan Genocide. According to the Financial Times, “[t]he bank is accused of transferring more than $1.3m of funds used to finance the purchase of 80 tonnes of weapons by a Rwandan general that allegedly broke a UN embargo and helped to arm the perpetrators of the country’s genocide.” Because much of my scholarly engagement with international criminal justice has focused on these sorts of allegations, for better or worse, I tend to assess normative accounts of the field with cases like this in mind. In what follows, I use the BNP Paribas case as a theme in commenting on Colleen Murphy’s excellent book, entitled The Philosophical Foundations of Transitional Justice (CUP, 2017). In part, I use this frame to shed light on an element of atrocity’s etiology that, in popular imagination at least, still seems shrouded in denial. More importantly for present purposes, though, I wonder if cases involving business participation in atrocity might offer a slightly different vantage point from which to evaluate conceptual models of transitional justice.

Murphy’s insightful account of the field points to four different features of societies in transition that make justice claims in these contexts a breed apart from those that apply in stable democracies. Contrary to Posner and Vermeule, Murphy eloquently argues that transitional justice is not like normal justice; it is qualitatively different because of the fundamentally distinct circumstances that constitute periods of political transition. As she explores in great depth, these surrounding political circumstances include: (a) pervasive structural inequality; (b) normalized collective and political wrongdoing; (c) serious existential uncertainty; and (d) fundamental uncertainty about authority. The combination of these circumstances in transitional contexts means that ordinary ideas of retribution, corrective justice, and restitution do not cater to the needs of affected societies. The misfit with these usual concepts of justice is so acute that we must craft an entirely new, distinct brand of justice that follows its own logic and answers to its own self-consciously defined first principles.

There is much I am attracted to in this approach and its brilliant execution. Let me start by pointing to the project’s apparent grounding in philosophical pragmatism. Surely, concepts of justice are context-sensitive. Undoubtedly, a major part of the problem for the field lies in avoiding blunt universalisms that do further violence to societies struggling to shed the heavy yoke of mass violence. At the same time, Murphy is clearly aware of the dangers of venerating the local at all costs; she astutely highlights the difficulties with over-correcting in favor of the local where she observes that “[r]adical contextualism can be just as problematic as an un-nuanced universalism.” (p. 81). Thus, transitional justice is philosophically pragmatic (by the by, I use this tradition to think about corporate responsibility here). Yet, if the concept of justice in transitional justice will always be context-sensitive, this leaves me wondering about the borders between the local and universal in Murphy’s self-contained concept of transitional justice, how this compartmentalized approach to justice will interface with others in a global society, and whether modulations in type of actor will affect the theory.

To begin, I am not entirely clear whether the BNP Paribas case falls within the bounds of transitional justice, on Murphy’s thoughtful definition or any other. On the one hand, it makes sense to treat the case as an element of transitional justice, assessing it against the very insightful standards Murphy has articulated. Apart from the alleged causal link to the genocide, our ability to alleviate the four characteristics that are particular to transitional societies is probably heavily dependent on transforming business—absent global economic revolution, post-conflict societies desperately need direct foreign investment that makes its way to people in survival economies instead of fuelling kleptocratic governments, corrupt patronage networks and brutal armed groups. On the other hand though, foreign economic actors are not clearly part and parcel of the affected society, precisely because they are outsiders culturally and geographically. The primary community of interest in discussing the moral legitimacy of the BNP Paribas case is likely French, not Rwandan, so the conceptual standards we would employ to scrutinize the case’s conceptual propriety lie beyond the special realm of transitional justice.

If the borders of a siloed concept of transitional justice are slightly unclear, the foregoing leads me to wonder about whether there is a geographical element to this theory. Are the principles of justice that govern transitional justice necessarily linked to features of a discrete, local political order? In other words, Murphy’s argument skillfully implies a plurality of justice (not just law), but I wonder if the different compartments of justice we create by disaggregating the concept aren’t overly geographically bound. We rightly think of Gacaca, the Rwanda Tribunal, and a host of other institutions and programs undertaken in Rwanda as falling within the transitional justice paradigm, in large part because the four characteristics that constitute the transitional paradigm stem from a single society reeling after mass violence within a single country. But importantly, the causal factors that gave rise to these atrocities were not geographically constrained in the same way. Foreign businesses in non-transitional societies allegedly contributed to the Rwandan Genocide too. So, which concept of justice—which set of principles—will govern the BNP Paribas case?

A causal account of atrocity might also suggest different rationale for justice. In her helpful discussion, for instance, Murphy points out that retribution does not easily match the needs of transitional societies since it assumes that wrongdoing is “deviant, individual and personal.” (p. 59). To focus just on deviance, the argument is that accountability is usually predicated on infrequent transgressions within an otherwise compliant society, so the retributive model collapses under the weight of widespread, normalized mass violence. Personally, I have always harbored vague misgivings about this argument, since it seems to assume a very temporally static, geographically bound notion of deviance (crimes in Nazi Germany were normalized at the time but deviant by even local standards measured before and after the war, and similarly, crimes in Rwanda were normalized there in 1994, but judged deviant by a thin concept of global community even at the time). Both these intuitions are very disputable, but there is another that business cases help bring into sharper relief—perhaps the actions of a particular foreign company implicated in atrocity were highly deviant even when violence was normalized for local citizens?

So, if the validity of justifications for punishment turn on which justice paradigm we choose and the type of actor implicated, these realities invite a wider set of reflections of how a compartmentalized concept of transitional justice interfaces with others in a globalized society. Here too, I wonder if the self-contained account is water-tight. For example, if globalization was born in colonialism, and actions of many foreign companies represent modern iterations of longstanding commercial practices consummated there, is there not a risk that a state-centric, geographically grounded concept of transitional justice will overlook the long history of foreign actors fomenting bloodshed for profit that at least dates to Leopold in the Congo? As for the present, what can we say about spikes in global demand for coltan for cellphones, tin for circuit boards and now cobalt for car batteries playing an important causal role in bringing about the mass violence transitional justice exists to remedy? Is there a danger that the very structure of transitional justice will leave these out?

Perhaps even the element of “transition” lends itself to this concern. Murphy rightly cites to a literature that queries whether transition matters to transitional justice, but cases like BNP Paribas add a new angle to those perspectives. In a conversation I had recently with the leading Argentine prosecutor of dictatorship-era crimes, he explained that as far as he could tell, the transition only implicated a particular cadre of political elites: “as for the businesses that supported the dictatorship, they remained the same before and after the transition.” If corporations are merely auxiliary to brutal authoritarian rule, this commercial continuity across political transition is probably less problematic, but if businesses enjoy a major share of power in affected communities, political regime change may not constitute a transition sufficient to move us into an analytically separate justice paradigm. Thus, for the businesspeople, it can be unclear what moral work the transition does in militating for or against their accountability.

I hope some of the foregoing provides useful food for thought, albeit from a relatively non-traditional vantage point. Without doubt, Professor Murphy’s rich and important book will animate discussions such as these for decades to come.

Defining the Field and its Moral Challenge


Laurel E. Fletcher is Clinical Professor of Law at UC Berkeley, School of Law where she directs the International Human Rights Law Clinic. Fletcher is active in the areas of human rights, humanitarian law, international criminal justice, and transitional justice. As director of the International Human Rights Law Clinic, she utilizes an interdisciplinary, problem-based approach to human rights research, advocacy, and policy. 


Colleen Murphy’s book “The Conceptual Foundations of Transitional Justice,” offers firm intellectual footing to transitional justice, a field that suffers from contested assumptions about what underlies its essential goals and methods. The question is will the book succeed in inducing practitioners, policy makers, and scholars to adopt Murphy’s theory of transitional justice, a moral theory based on relational transformation? Such adoption would in turn lead to different approaches to transitional justice and challenge our current strategies.

The United Nations has legitimated and defined transitional justice as “the full range of processes and mechanisms associated with a society’s attempts to come to terms with a legacy of large-scale past abuses, in order to ensure accountability, serve justice and achieve reconciliation.”[1] However, this broad definition tells us little about what specific outcomes these interventions are supposed to achieve: what is accountability, justice, or reconciliation? Moreover, what are the principles that should guide interventions, and how we can discern whether we are getting closer to a desired end state? These shortcomings are symptoms of a field plagued by under-theorization and no agreed-upon theoretical foundation on which to develop a transitional justice framework.

The lack of theoretical grounding can be traced, in part, to the manner in which transitional justice developed. The field grew as primarily a legal response to the political question of how new governments succeeding repressive or authoritarian regimes in Latin America and eastern Europe, respectively, should address the mass human rights violations committed by a prior regime? Geopolitical trends – the end of the Cold War – opened up space for human rights claims of accountability for perpetrators and justice for victims to be realized. In remarkably short order, “accountability” and “justice” displaced amnesty for dictators and warlords and became the politically hegemonic defaults for addressing mass abuses.

Leveraging transitional justice as a rallying cry, advocates across diverse and varied contexts demanded action. The frame of transitional justice united calls in Spain to provide justice to victims of political violence of the Franco regime with demands to initiate international prosecutions during active conflicts in the Great Lakes region with grassroots efforts to promote reconciliation for racial violence in the United States. When I served as an Editor-in-Chief of the International Journal of Transitional Justice, I was struck continually by the diverse cases that authors argued qualified as falling within the rubric of transitional justice. But the capacious UN definition of transitional justice, while globalizing the field and swelling its ranks of adherents, creates other problems. Can what we consider to be appropriate responses to egregious violations that arise in vastly different contexts be universalized in any meaningful way?

Murphy makes the case that the answer to this question is a qualified “yes.” Approaching the question from the perspective of moral philosophy, Murphy investigates how to evaluate the justness of the legal responses that societies make to respond to past violence and repression. She asserts that the question of whether a response is “just” is context-specific, and therefore in order to answer what justice looks like in transitional contexts, we must first know what demands transitional justice needs to satisfy. In transitional settings, according to Murphy, the central moral problem that justice must address is societal transformation. Societal relationships have been distorted by pervasive structural inequality and normalized collective wrongdoing. The correct moral response, she argues, is to reestablish reciprocal relations between state officials and citizens based on fulfilling mutual duties based on respect for human agency and rule of law. Thus, she argues, her positive moral theory of transitional justice is capable of guiding legal responses to achieve the unique justice requirement in transitions.

Because her moral theory is premised on the assumption that transitional justice is distinct from other forms of justice, Murphy develops a model to define the characteristics, and therefore the parameters, of transitional justice. She argues that transitional justice is a morally required response to situations in which four conditions are satisfied: (1) pervasive structural inequality exists; (2) normalized collective and political wrongdoing occurred; (3) there is serious existential uncertainty that a transition to democracy will occur; and (4) there is uncertainty about the trustworthiness of authority, e.g. the state is complicit in the wrongdoing. Under these conditions, Murphy argues that conventional theories of justice do not apply because retributive, restorative, and distributional theories of justice assume a background of a stable democracies. In stable democracies, the fundamental problem justice for a single murder must address is retribution for the perpetrator. In transitional justice, for example where the State committed genocide, the fundamental justice challenge is not delivering just desserts to wrongdoers but transforming society.

In developing her theory, Murphy addresses several preoccupations of transitional justice scholars and practitioners. First, there is a widespread understanding among adherents that transitional justice responses have to include retrospective and prospective dimensions. Accountability for perpetrators of widespread and severe wrongdoing is needed. But to stop abuses from recurring, societies need to solidify rule of law based on respect for human dignity. Second, there is an increasing awareness that structural marginalization and discrimination are drivers of mass violence and abuses and therefore post-conflict responses must include rectifying these problems. Third, there is an emerging consensus that transitional justice requires a holistic approach. No single mechanism or program will be sufficient to resolve the myriad harms caused by past episodes of widespread violations. Murphy unifies these concerns into a single moral frame in which legal interventions are directed at societal transformation.

In synthesizing these strands, Murphy makes an important contribution. By reframing justice away from legal accountability and toward the evaluation of legal responses based on their contributions to reforming political relationships, both between citizen and state and among citizens, she moves past the increasingly unsolvable debates that have preoccupied the field. These debates have focused loosely around two sets of questions. One set of questions concerns the appropriate balance among the legal duties to accountability, truth, reparations, and measures of non-recurrence. The second set of questions focuses on the value of any particular perspective—law, culture, power—to guide thinking about the goals, methods, and processes of transitional justice. Neither set of issues is settled and often there is cross-talk between these sets of questions, which confuses matters further.

In other words, Murphy’s argument switches our attention. If the central goal of transitional justice is to transform political relationships, we ask different questions of legal interventions than if we assume that the “justice” of transitional justice is equated with retributive justice criminal prosecutions, or restorative justice or truth telling, etc. If we focus on realizing the moral commitment to achieve relational transformation we might ask instead: What might it take to restore public trust, establish reciprocity and respect of individual agency between state officials and citizens? The answers might include legal measures but would likely extend far beyond.

Her case studies of Uganda and the former Yugoslavia illustrate the conceptual blind spots of conventional forms of transitional justice to promote societal transformation. Justice in these cases was not just because legal interventions did not attend to the moral relationship of the State to victims, and to perpetrators. But, as she admits, her moral theory supplies conceptual tools for considering and evaluating legal interventions rather than a blueprint for policy prescriptions.

Nevertheless, Murphy’s moral theory does have practical implications. It demarcates transitional justice cases from other cases of injustice that deserve attention. Not every case of mass violence that has been incompletely addressed in the eyes of victims can be converted into a case requiring “transitional justice” by claiming that society has failed to reckon with its past. For example, cases of historical injustice in stable democratic regimes (e.g. abuses under the Franco regime or slavery in the United States) fall outside of Murphy’s theory because the state is no longer “in transition.” Democracy is in place. There may still be lingering ‘injustices’ but these do not require societal transformation of relationships that Murphy’s moral theory requires.

Defining parameters for transitional justice is courageous because it narrows the field and the number of stakeholders invested in it. Another ramification is to discipline the enterprise of transitional justice by establishing parameters around what values are legitimate to pursue and through what methods. If Murphy’s theory is adopted, it would reorient the field away from questions of legal accountability and toward a focus on questions of the nature of moral relationships between citizens and the state.

Without a firm grounding in theory, transitional justice risks becoming a slogan to be manipulated by any number of actors. To defend the field, we need to know what is unique about society’s need to respond to mass violence and to refine what form of justice society is called up to do in those instances. Whether Murphy’s argument succeeds to reshape the field is unclear. However, her argument invites and deserves important debate. The Conceptual Foundations of Transitional Justice deserves to be read.

[1] Report of the Secretary-General, The Rule of Law and Transitional Justice in Conflict and Post-Conflict Societies, 23 Aug. 2004, S/2004/616, Para 8.

Practical Reflections on Conceptual Foundations


David Tolbert was appointed president of the International Center for Transitional Justice in March of 2010. Previously he served as registrar (assistant secretary-general) of the Special Tribunal for Lebanon and prior to that was assistant secretary-general and special expert to the United Nations secretary-general on United Nations Assistance to the Khmer Rouge Trials.

Roger Duthie is Director of Research at ICTJ, where he has managed research projects examining how transitional justice relates to education, forced displacement, and development. His publications include Justice Mosaics: How Context Shapes Transitional Justice in Fractured Societies (2017, co-edited with Paul Seils); Transitional Justice and Education: Learning Peace (2016, co-edited with Clara Ramirez-Barat); Transitional Justice and Displacement (2012); and Transitional Justice and Development (2009, co-edited with Pablo de Greiff); as well as articles published in the International Journal of Transitional Justice and the International Human Rights Law Review. 


In The Conceptual Foundations of Transitional Justice, Colleen Murphy develops a theoretical framework for understanding the conditions, objectives, and processes of transitional justice. It is a very interesting and useful contribution to the literature on transitional justice. The author’s approach is to initially take a step back from questions of transitional justice measures and processes (the focus of most practitioners) and pose more fundamental inquiries: Under what conditions do the need for transitional justice arise? What problem or issues is transitional justice addressing? How does it respond to that problem? In doing so, the book helps us to rethink how we—practitioners, interested parties, and scholars—can more coherently, effectively, and justly respond to past wrongdoing.

Murphy posits transitional justice as a singular or separate type of justice. She argues persuasively that transitional justice is not simply a refined version or combination of retributive, corrective, and distributive justice, but is a different kind of justice, which focuses on specific problems and offers normative guidance on how to respond to those problems. Central to Murphy’s understanding is the transitional nature of the context in which societies address wrongdoing, which she finds as necessary for defining transitional justice. In her view, it is not the achievement of transition that is needed, but rather the aspiration to end conflict and replace repression with democracy; at the same time, however, it is also Murphy’s view that not all cases commonly labeled “transitional justice” should in fact be included in the category. While positing democracy as the end goal of transitional justice is widely shared by victims and practitioners alike, this is often a long-term goal. Moreover, the content and understanding of democracy varies widely.

The book identifies four conditions or circumstances that are necessary to give rise to the specific problem that transitional justice is to address. The first condition is pervasive structural inequality, which refers to the illegitimacy of the institutional rules and norms shaping interaction among citizens and between citizens and officials. The second is normalized collective and political wrongdoing, in which violations of human rights have become basic facts of life through different forms such as centralized repression or symmetrical or unstructured violence. The third and fourth conditions are serious existential uncertainty, referring to the very unclear trajectory of a political society, and fundamental uncertainty about authority, or the standing of a political regime to rule and enforce rules.

There are two particularly important points about these conditions for the book’s argument. First, the different conditions affect each other: pervasive structural inequality, in particular, is empirically correlated and mutually reinforcing with normalized wrongdoing, and it is also necessary for existential uncertainty. “Particular incidents of violence are part of a broader pattern of interaction designed to entrench and reinforce pervasive structural inequality,” Murphy writes. “Such wrongdoing needs to be seen against the background of pervasive structural inequality, and as intended to reinforce and be justified by that inequality” (page 103). As she put is, “background injustice is the subject of transitional justice” (page 95). This means that it is not reform of institutions and norms that is needed, but transformation. Second, the uncertainty about the nature and direction of a political order highlights the importance of identifying how different factors may affect outcomes: how does one determine how to bring about the transformation that is needed?

For Murphy, then, the core moral question for transitional justice is how to justly pursue societal transformation. How does a society transform the structure of political relationships so that it is based on respect for agency and reciprocity? As she notes, in this context transitional justice can be linked to reconciliation, or more aptly improving damaged relationships, a subject on which she has written a separate book. In this book, Murphy argues that such a transformation depends on establishing respect for rule of law, a certain extent of relational capabilities, and reasonable political trust, and that transitional justice can contribute to this transformation in direct and indirect ways—for example, by acknowledging the need for change and giving people hope that such change can in fact come about.

The book also contends that specific transitional justice processes should constitute fitting and appropriate responses to victims and perpetrators of past wrongdoing. The processes should, in other words, respect the core moral claims associated with being a victim or a perpetrator of wrongdoing. This, she suggests, can be assessed by taking into consideration the moral aims of the response, including repudiation, accountability, acknowledgment, recognition, reparation, and non-recurrence; the relationship between the responder and the subject of the response; the nature of the wrongdoing, in that it should deal with actual harms and respect due process; and the cultural norms and consequences of the action. Furthermore, responses to wrongdoing should be holistic, in that they should be multiple and coordinated, because, among other reasons, no single response can achieve the relevant moral aims, while the expressive meaning of each response is shaped by other responses.

The short case studies in the final chapter focus on the Ugandan amnesty commission and the International Criminal Tribunal for the former Yugoslavia. If the point of these examples is to show the inherent limitations of single responses, however, these seem unusual choices, as an amnesty body and an international tribunal bring with them a whole set of issues that most national-level transitional justice processes do not necessarily have to deal with. Furthermore, while not disagreeing with Murphy’s assessment of the flaws of these processes, it is important to note that some flaws are inherent limitations of the type of response, and some flaws are correctable or avoidable process issues. Moreover, given the limitations of each specific process, they will only address certain issues; thus, the question may be one of overall design or political will rather than the mechanisms themselves, which are established for particular purposes. As she points out, no program of reparations in Uganda could have dealt with the layered harms and social consequences experienced by victims, for example, but courts can take steps to be less removed from local populations.

Interestingly, the book explicitly does not examine in detail specific responses to wrongdoing, or the contribution that such responses may make to societal transformation or the moral claims of victims and perpetrators. As the author explains in her conclusion that this is because she believes our understanding of the moral functions, impact, and expressive meaning of such responses is limited, and that more research is needed before such a discussion should be had. Furthermore, with this book Murphy writes that she wants to encourage societies to expand the range of responses to wrongdoing that they consider, and discussing specific existing responses—which, it is true, tend to cluster around criminal prosecutions, truth telling, reparations, and certain kinds of institutional reform—may reinforce what exists now as the full range of options. She points to areas such as art, theater, and television, debt forgiveness and land redistribution, and education as potentially important responses to wrongdoing that may fall outside the current set of measures generally considered transitional justice.

In this, she echoes a 2015 report of Pablo de Greiff, UN Special Rapporteur on truth, justice, reparation, and guarantees of non-recurrence, emphasizing the relevance of non-institutional responses in the realm of the cultural and the individual in preventing the recurrence of violations.[1] Also relevant is the notion of “transformative justice,” which some argue would target the structural causes of human rights violations more directly than transitional justice. Murphy does not address this notion explicitly, but she is similarly arguing for an expanded response to broad injustice, although it seems within the concept of transitional justice rather than within a new notion. Either way, the argument raises significant questions about the distinction and relationship between responses to human rights violations and processes such as development and peacebuilding. As the book notes, ICTJ has conducted research on education as an important part of the way in which societies respond to wrongdoing, although without necessarily explicitly categorizing educational initiatives as transitional justice.

The book also makes an interesting argument regarding the requirements for transitional justice to be considered just. Drawing on the structure of just war theory, which requires that armed conflict satisfy both requirements of jus ad bellum and jus in bello, Murphy argues that in order to be morally defensible, responses to past wrongdoing must satisfy two requirements: a morally permissible objective, in that they must contribute to societal transformation; and a morally permissible manner, in that they must be fitting and appropriate, as discussed above. She acknowledges that the just nature of different responses is scalar—that is, they can be assessed as being more or less just.

Murphy’s argument that pervasive structural inequality is so integrally connected to normalized wrongdoing is persuasive that transforming political relationships should be an equally integral element of transitional justice. However, the claim that responses to wrongdoing necessarily fail to be just if they do not contribute to societal transformation seems debatable and run counter to some of Murphy’s own argumentation. As she acknowledges, to see transitional justice as only instrumental is problematic: claims for transitional justice “have an independent, non-instrumental moral importance. To regard wrongdoing as valuable only instrumentally is insulting to victims and insufficiently respectful of the agency of perpetrators” (page 114). If responses to wrongdoing that respect the moral claims of victims and perpetrators have an independent moral importance and therefore should not be seen as having only instrumental value, why then must they also be instrumental in order to be just? One can argue that responses that are not instrumental to societal transformation may not qualify as transitional justice, according to Murphy’s conceptualization of the notion, but they can still be morally defensible and just, can they not?

The book also makes a compelling case that democratization is a necessary part of the societal transformation to which transitional justice can contribute. This will be contested by others, but it seems reasonable to argue that democracy is necessary for the establishment of political relationships among equals based on the values of reciprocal agency, which is necessary in turn for the protection of fundamental human rights and the prevention of the recurrence of normalized wrongdoing. One can ask, however, that if democracy is necessary to societal transformation, then why is peace not? Is peace not necessary for the protection of fundamental rights? Murphy does discuss transitions out of conflict, and specifically notes that the cessation of violence can contribute to giving hope to people that change is possible. But she does not explicitly make the case that the prevention of the recurrence of armed conflict should be considered a necessary part of societal transformation from a justice perspective.

While the book does focus on conceptual and, in some sense, scholarly issues, it also will be useful to practitioners. Many of the issues that have been surfaced by Murphy are ones that we, as practitioners, face on the ground. For example, issues of marginalization and of economic exclusion and of the goal of democratization are important elements in our work. In some cases, they are not thought through as thoroughly as we would like. This book helps provide a framework for those kinds of discussions, which helps practitioners find practical solutions.

[1] UN General Assembly, Report of the Special Rapporteur on the promotion of truth, justice, reparation and guarantees of non-recurrence, Pablo de Greiff, UN Doc. A/HRC/30/42, September 7, 2015.

New Symposium: The Conceptual Foundations of Transitional Justice


Colleen Murphy has written an excellent and important book, entitled The Conceptual Foundations of Transitional Justice (CUP, 2017), that a distinguished set of scholars from a range of disciplines begins to discuss. The rise of Transitional Justice as a distinct field over the past several decades has enjoyed much excellent literature, but Colleen Murphy’s new book is among the first to offer a dense philosophical account of the field. Necessarily, this account will hold much of interest to scholars from philosophy, but it will also have implications for international lawyers, criminal law theorists, international criminal lawyers and human rights scholars. This blog also seeks to showcase groundbreaking, normatively creative new works that reach a broad array of scholars and practitioners alike, and this book certainly meets that mark. The Conceptual Foundations of Transitional Justice will, no doubt, be a major contribution to the field and discussed for decades to come, so it is a great pleasure to begin those conversations with an expert cast of highly accomplished scholars and leading practitioners.  

I will not provide a detailed overview of the book except to highlight important themes.  In Chapter 1, Murphy starts be plotting the circumstances of transitional justice, which she argues demonstrate four elements. Justice in transitional societies is different in type because these societies experience all of the following: (a) pervasive structural inequality; (b) normalized collective and political wrongdoing; (c) serious existential uncertainty; and (d) fundamental uncertainty about authority. In Chapter 2, Murphy argues that ordinary ideas about retribution, corrective justice, and restitution do not cater to the needs of societies undergoing transition according to her definition. These segments of the book are richly informative about both the political context of transitional society, and their intersection with philosophical bases for different theories of justice. The resulting account rejects Posner and Vermeule’s notion that Transitional Justice is just like ordinary justice, positing the existence of a conceptually self-contained set of principles governing the field. In Chapter 3, Murphy provides a substantive account of societal transformation, drawing on relational transformation, rule of law and relational capacities. In Chapter 4, she highlights constraints on this transformation, based on whether Transitional Justice initiatives are “intrinsically fitting or appropriate as a response to victims or perpetrations of wrongdoing.”  

I am excited to host a range of leading experts from a variety of fields.  Roger Duthie and David Tolbert are the Director of Research and President of the International Center for Transitional Justice respectively, an organization which works with victims, civil society, and national and international organizations within countries that have endured massive human rights abuses. Duthie and Tolbert provide fascinating commentary on the book through the lens of the ICTJ’s active work.  Laurel Fletcher directs the International Human Rights Law Clinic at UC Berkeley, School of Law, adopting an interdisciplinary, practical approach to human rights issues.  Fletcher’s contribution speaks to the lack of theoretical grounding in the field of transitional justice up until now, and what impact the adoption of a moral philosophy might have.  The work of Nir Eisikovits of University of Massachusetts Boston focuses on philosophy and applied ethics, in particular the ethics of war and political philosophy.  He describes Murphy’s book as “the best, most ambitious philosophical account of transitional justice” he has ever read and points to important implications arising from it. And last but not least, Steven Ratner at University of Michigan Law School focuses on Murphy’s call for “an original methodological hook for analyzing the justice of a state’s transitional justice choices,” exploring the legal ramifications of Murphy’s theories.  My own comments offer thoughts about compartmentalizing transitional justice, using cases against foreign businesses as a vantage point.  

As an ensemble, I hope the resulting body of perspectives showcases this important work and offers fresh ideas for future thinking.  

An Open Invitation to Further Debate (Instead of an Amicus Brief)


In organizing this mini-symposium, I sought to engage expert reactions to my paper from a range of legal systems that have not featured in debates about forms of attribution in ICL. When international courts and tribunals construct(ed) these forms of attribution, they initially drew heavily on Anglo-American jurisdictions, adopting concepts like superior responsibility and joint criminal enterprise from them. Then, judges at the ICC announced a major swing towards notions of criminal responsibility derived from German criminal law, including co-perpetration, theories of control to distinguish perpetration from complicity, indirect co-perpetration and even perpetration through an organization to treat those doing the bloodletting and their masterminds as perpetrators. Throughout this process, nobody appears to have asked experts in systems that adopt a unitary theory of perpetration, which dispenses with all these doctrines, to reflect on the law within their own countries and its potential as a solution to recurrent problems with blame attribution in ICL. This silence has been quite strange, especially when the Nuremberg Tribunal applied a unitary theory of perpetration and several modern ICL judges have argued that the current complexity is unnecessary.

This mini-symposium has broken new ground in this regard, in ways that I hope sets the scene for further scholarly research and debate. I was especially grateful that a range of criminal law theorists from each of the countries I write about in the paper agreed to criticize the paper, and that some very prominent practitioners joined the fray to offer their reflections too. As is evident from this blog’s manifesto, I deliberately seek to create dialogue between theorists and practitioners, so I am thrilled that this discussion has involved members of both groups. Some of the feedback I received was striking—during the course of this online symposium, a senior prosecutor at one international court and a defense counsel for a well-known defendant at another emailed me saying they wholeheartedly agree with the need for a unitary theory. The latter even suggested that I file an amicus brief calling on one particular tribunal to revert to the unitary theory of perpetration adopted at Nuremberg. I politely declined, but decided to open up this final post to whomever wanted to share an opinion one way or the other, provided that it respected the strictures of the blog’s manifesto.

Instead of defending the unitary theory of perpetration or either of the article’s I’ve written about it (see here and here) in this post, I use this opportunity to set the scene for an open online discussion at the base of this post by reiterating what a unitary theory is and by summarizing the excellent posts that appeared in this symposium.

To begin, let me again highlight the three main variations of the unitary theory of perpetration to avoid commentators speaking past one another. The unitary theory of perpetration comes in three principal varieties, although some might contest whether the third species really fits within the genus. The first, known as a pure unitary theory, treats a causal contribution to a crime coupled with the requisite blameworthy moral choice announced in the criminal offence charges as necessary and sufficient elements of responsibility (excuses and justifications aside). On this view, the various forms of participation that exist in current ICL (aiding and abetting, JCE, co-perpetration etc.) are stripped of their autonomous existence and folded into a more capacious single notion of attribution. So, instead of attempting to manufacture fine-tuned rules that define JCE, aiding or any other form of participation in such and such a manner, a unitary theory of perpetration places them all in a big pot, then boils them all down to their shared normative essence. Through this distillation, blame attribution involves deciding whether accused X is responsible for crime Y based on settled core principles that pay no regard to the form participation takes, leaving their moral significance to be assessed post hoc by judges at the sentencing phase of a trial.

The second variant provides more detail without compromising the unitary theory’s core commitments. Known as a functional unitary theory, this iteration provides more guidance while insisting that causation and the mental elements announced within the criminal offense charged are necessary and sufficient bases for establishing wrongdoing across all forms of participation. To ensure that would-be criminals are sufficiently forewarned of their exposure to criminal law penalties, a number of states adopt this variant of a unitary concept—the general part of a criminal code or legislation articulates the different forms of causal connections that might apply within a unitary framework. In this sense, responsibility might involve carrying out the offence personally, instructing others to do so, providing necessary assistance, or furnishing assistance that is readily available elsewhere. Each of these forms of causation is announced within the law so as to inform the public of how they might attract criminal responsibility, but the underlying objective and subjective elements beneath these descriptions remain the same.

Third, some argue that subjecting accomplices to the same range of punishment as perpetrators also constitutes a weak type of unitary theory. In Germany (and the many jurisdictions that follow its example), aiders and abettors are sentenced to a maximum of three quarters of the penalty for the offense they facilitate, whereas the sentence for instigators is taken from the same sentencing range as principals. To a large extent, this discrepancy in maximum sentence drives the need for differentiating between perpetrators and accomplices, even if, as Markus Dubber has observed, “[r]emarkably little effort is spent on justifying this differentiation”.[1] Nonetheless, this differentiated approach, whose purpose is partly to determine the applicable range of sentencing, generates a tendency to look upon systems that formally equate sentencing ranges for perpetrators and accomplices as soft iterations of the unitary theory. France and England, for instance, do just this. For my purposes, though, I do not consider this an example of the unitary theory because it places no restriction on the substantive elements of forms of attribution, whereas truly unitary theories do.

With the stage set, I next situate the various expert responses to this mini-symposium, grouping them into those who also advocated for a unitary theory in ICL, those who were more ambivalent about whether their national experience served as much of a template for ICL, and one who was positively unconvinced.

In the first of these categories, Judge Baragwanath’s excellent post reminded us that there are actually many jurisdictions that fit within variants two and three, even if they might not describe themselves as unitary theories of perpetration. My own country of origin, New Zealand, begins the provision governing parties to offences by stipulating that “[e]very one is a party to and guilty of an offence who,” before articulating different forms of participating in a consummated offence. Judge Baragwanath’s post is so useful because it not only highlights that New Zealand’s criminal law is, in important aspects, unitary, but it also shows how a series of cases in England, Australia and Hong Kong have been struggling with whether to tie mental elements in forms of participation to those in the offense announced in ways that mimic the unitary theory. Despite backsliding in some courts, there is a discernable modern trend in this direction. His post reminds me that the States of New York and California have an even more intense unitary theory of perpetration. In any event, in describing “modes of liability” as “unnecessary,” Judge Baragwanath argues that “international criminal procedure, already complex and expensive, adds to those problems by forcing itself to leap over self-created non-existent hurdles.”

Filippo de Minicis’ post is similarly minded. Filippo is a presently Legal Officer in the Office of the Co-Investigating Judges in the Extraordinary Chambers in the Courts of Cambodia, but he was originally trained in Italian criminal law, which as I show in the article, also discarded a differentiated system of blame attribution in favor of a unitary alternative almost a century ago. Filippo argues that when looking at standards of attribution before ad hoc international criminal tribunals (i.e. in customary international law), there is “little difference in the required actus reus,” and “a sufficient homogeneity on the mens rea side.” Filippo concludes after a decade working with these standards that a unitary theory is both viable and preferable, but he is also circumspect about whether any theory is perfect and, as was the case with New Zealand, shows how Italy’s commitment to the unitary theory is not absolute since Italy too appends a type of common purpose doctrine that approximates to JCE. Reality, it turns out, is complicated. Despite this, Filippo joins Judge Baragwanath as an advocate for the adoption of the unitary theory in ICL.

Other commentators are more ambivalent. Professor Carlos Eduardo A. Japiassú, for instance, highlights how Brazil’s unitary theory has slowly changed over time, shifting from a pure version to a functional one (which he calls “mixed”). While Professor Japiassú also speaks of a certain legal conservativism in Brazil, which I take to imply a lack of desire to shift back to a differentiated system that employs different substantive tests for different stand-alone forms of participation, he ends by concluding that “it remains unclear whether a pure rendition of this theory or a mixed variant like that now applicable in Brazilian Penal Law is a good alternative for International Criminal Law or International Criminal Courts.” Similarly, Professors Iryna Marchuk and Jørn Jacobsen discuss important scholarly criticism of the unitary theory in Denmark and Norway respectively as well as partial retreats from it in recent doctrine, before also questioning their system’s value as an exemplar for ICL.

Finally, in the third category, Judge Albin Eser’s masterful critique exemplifies disagreement with the unitary theory. In many respects, his is a brilliantly concise defense of the structure of blame attribution currently in place in ICL and a deft rebuttal of the arguments in the paper. The series of questions he poses are skillfully listed as issues he would need to be convinced of to accept that a unitary theory is optimal. These start with the argument that different forms of participation actually better track real life, move to the idea that a unitary theory cannot justify why they are addressed at sentencing along, then shows how unitary theorist essentially overlook that these questions will arise at sentencing anyhow. Then, he argues that “the only practical advantage the unitary theory so far seems to offer is a procedural one,” but he sees no procedural advantage here either. Ultimately, he concludes by correctly pointing out that even if we do have a differentiated system of blame attribution in ICL because powerful western states forced it on others, this says nothing about the theory’s conceptual integrity. A unitary theorist would, of course, contest each of these steps, but Eser’s brilliant critique is a wonderful counterpoint.

So, instead of labouring my own perspective any further here, I make space for other scholars, experts and practitioners to weigh in on these debates, which strikes me as a better idea than filing an amicus brief. I have therefore opened this post to comments, and anyone can post their views directly. In order to help ICL practitioners share their views (I recall many hours debating these questions with colleagues in war crimes tribunals), I’d like to offer a procedure through which you can legitimately (I hope) bypass the need for institutional approval to publish. If your institution is agreeable, I will post thoughts and reflections from practitioners anonymously. I would not normally do this through the post, so if want to remain anonymous, please send me your comments by email at stewart@law.ubc.ca. Your email message to me should include your title and the institution you work for, but I undertake to keep this information entirely confidential, posting only your thoughts and reactions on this topic. For the rest of you, the post is open.

[1] Markus D. Dubber, ‘Criminalizing Complicity: A Comparative Analysis’, (2007) 5 Journal Int Criminal Justice 984 ff.

Questions From the Unconvinced


Albin Eser is Director Emeritus and Professor Emeritus of Criminal Law, Criminal Procedure, and Comparative Criminal Law with the Law Faculty of the University of Freiburg.  He was formerly a Judge at the International Criminal Tribunal for the former Yugoslavia.


To be satisfied with the unitary theory, you must be content with the most simple solution. In my view this is not the best approach to cope with the variety of social and criminal life. As soon, however, as you wish to pay attention to the varieties of the performances of crimes, you must be prepared to make differentiations – and this may unavoidably be a source for different demarcations. But even if this entails controversies, can this be a reason to sacrifice individual justice (by distinguishing between perpetration and participation) on the altar of (unitary) simplicity?

To become convinced of the latter alternative, I would like to have these questions answered:

  • Regarding the argument that the differentiation models would not correspond to real life, is this more the case if causation is considered the only basis of imputation, thereby ignoring any differences in the manner in which, and in the weight by which, a contribution is made to the performance of a crime?
  • And if there is a factual and social difference, as can hardly be denied if the view is not normatively and holistically preprogrammed, should this difference be paid attention to only at the sentencing stage or should it rather be expressed already in the guilty verdict?
  • And if only on the sentencing level, how and according to which criteria should this be done? This is a question, by the way, which is neither dealt with in your article nor are any suggestions by your proponents of the unitary theory visible.
  • If indeed, however, even the unitary theory cannot avoid taking notice of different types and manners in which a person can be involved in the performance of a crime, why couldn’t and shouldn’t this be made public in the verdict?
  • So the only practical advantage the unitary theory seems to offer so far is a procedural one: that the problem of recharacterization may be avoided. Yet, is this really the case if even according to the unitary theory differences in the crime performance are (to be) made at the sentencing stage? In a murder case, for instance, what could a defendant do if he had defended himself by merely involuntary having told where the victim might be met but finds himself sentenced to life imprisonment because of being proven to have been on the scene and directly involved in the killing? Could his appeal be rejected by simply referring to his causal contribution, thus rendering any differential circumstances irrelevant? Or shouldn’t it rather be in the interest of individual justice to assess the penalty according to the type and weight of his contribution? However, if he thus succeeded in getting his sentence adjusted to his minor contribution, as I think he should, could this be done without differentiations and recharacterizations? So what, in the end, is finally left of assumed procedural advantages of the unitary theory?
  • Regarding your references to political and ideological superpowers smaller countries were able to free themselves from by introducing the unitary theory, should such sovereignty aspects indeed play a role as to whether criminal theory is good or bad?

A Unitary Theory is Both Viable and Preferable


Filippo de Minicis works as a Legal Officer in the Office of the Co-Investigating Judges in the Extraordinary Chambers in the Courts of Cambodia. After graduating in law in 2004, he practiced criminal law in Italy. From 2006 to 2013, he worked first as a defence consultant and then as a Legal Officer in Chambers at the International Criminal Tribunal for the former Yugoslavia.


I agree with Professor Stewart’s proposition that a unitary model of blame attribution would be a viable and preferable option for ICL. This opinion is inevitably rooted, at least in part, in my Italian legal training. It is, however, also based on my ten years’ experience as a practitioner in ICL. Because my professional experience is limited to Italy, the ICTY, and the ECCC, my focus will be on participation as applied in those fora.[1] My observations will cover three areas: 1) the issue of the Fascist origin of the Italian Penal Code; 2) the reasons why I favour a unitary system to the current model; and 3) one aspect of the Italian system that Professor Stewart will probably not like.

 1.  Codice Rocco – An essentially liberal code in Fascist attire

Professor Stewart notes that the Fascist origins of the Italian Penal Code (known as Codice Rocco) have led some to dismiss it as a possible model for a unitary theory of participation. Stewart rejects this criticism. I do too. The Codice Rocco is divided in two parts: a general part, which contains the general principles of criminal law, such as causation, culpability, and participation; and a special part, which defines crimes and prescribes penalties. It is in the latter that the Fascist footprint is most evident, both for the nature of certain crimes (e.g. article 265 of the Code prohibits “political defeatism”) and for the order in which they appear, which echoes the hierarchy of values of the Fascist regime. The general part, however, opens by introducing the principle of legality, the prohibition of retroactive application criminal law, and the prohibition on recourse to analogy as fundamental principles of criminal law. These are principles that predate the Fascist ideology and which were part of the legal education of those who drafted the code. They stand in stark contrast to the general principles governing German criminal law during the National Socialist rule, which in substance left to the judge the freedom to incriminate whoever exhibited a behaviour that appeared to conflict with the interests of the National Socialist Party. The Codice Rocco is thus essentially a sufficiently liberal and still modern code, and while there are some articles of the general part which were influenced by the Fascist ideology, the unitary system is not one of them.

2.  A unitary model is possible and may have practical advantages

Italian law does not distinguish between principal perpetrators and accomplices. By contrast, both the ICTY and the ECCC adopt such a distinction. As an Italian lawyer trained to examine criminal liability through a monistic lens, I see the modes of liability applied by the ad hoc tribunals as an unnecessarily prescriptive characterization of different ways of participating in a crime. I hold this view because:

  • Apart from superior responsibility,[2] all modes require that the accused, by ordering, instigating, planning, or aiding and abetting,[3] substantially contributed to a crime, or that her conduct had a substantial effect on its commission.
  • Similarly, for Joint Criminal Enterprise (JCE), the law requires that the accused made a significant contribution to the common criminal plan.
  • With respect to the subjective element, they require that the accused made this contribution with the intent that the crime be perpetrated, or with the awareness of the substantial likelihood that a crime would be committed as a consequence of her conduct.
  • For aiding and abetting, the required mens rea is knowledge that the aider’s conduct will substantially contribute to the commission of the crime.
  • There is thus little difference in the required actus reus,[4] and a sufficient homogeneity on the mens rea side (knowledge, in my view, would also be a sufficient standard to attribute liability through a unitary model – see the next paragraph).

These all appear to be good reasons to consider whether the existence of a variety of modes of liability with an identical or sufficiently similar mens rea requirement is necessary. While some differences, especially if we consider the ICC Statute, do remain in the current universe of modes of liability internationally, they are of a nature that they can be reconciled in a unitary model, especially with appropriate adjustments in the sentencing law and practice. The means with which a person causes or contributes to a crime are not generally determinative of criminal liability strictu sensu. At most, they may be factors relevant for sentencing.

A word about the knowledge standard for aiding and abetting. Leaving aside Professor Stewart’s correct observations that some judgments of the ad hoc tribunals actually require a recklessness standard for aiding and abetting (i.e. knowledge of the probability that her conduct will assist the commission of the crime),[5] the knowledge standard as interpreted by the ad hoc tribunals would suffice, in Italy, to convict an accused for an intentional crime, i.e. one requiring dolus. By way of example, if I want to kill my business associate, and I ask a friend of mine to lend me his gun informing him that I need it to perpetrate the murder, if eventually I commit the murder both my friend and I will answer for it as perpetrators. I see no dogmatic problem with this approach: knowing that our conduct will cause or is likely to cause certain consequences, and still engaging in such conduct, essentially amounts to an intentional act, irrespective of whether those consequences were also desired.

In my experience, it is often practically difficult to distinguish whether a person contributed to a crime with knowledge (in the sense of awareness that one’s conduct will (likely) produce certain consequences) or intent (in the sense of knowledge with the added element of the desire to achieve certain consequences).[6] This, in turn, renders it problematic to legally characterize one’s responsibility as commission through a JCE or as aiding and abetting.[7] Let’s take the example of an alleged JCE the criminal objective of which is the deportation of an ethnic minority. To establish liability through JCE, the prosecution will need to prove that the accused, in concert with others, significantly contributed to the deportation with the intent to commit that crime. In practice, however, it may be difficult (or even impossible) to distinguish this situation from that of someone whose conduct (e.g. the provision of buses to deport the minority) substantially contributed to the deportation with knowledge that the provision of those means would assist that crime, but without the desire to achieve that result: in this scenario, the accused would incur liability via aiding and abetting. In ICL cases, there is seldom a “smoking gun” that allows a simple determination of one’s state of mind and, especially in JCE cases, intent is often inferred from a combination of knowledge and continued participation. This system, however, is far from infallible. Personally, I am not convinced that it is always (or even often) possible to distinguish if a person acted with mere knowledge or intent. Consequently, the law of the ad hoc tribunals features a theoretical distinction that the judges may not be able to apply – beyond reasonable doubt – to the facts under their scrutiny.

Thus, considering that all modes of liability (apart from – to some extent – superior responsibility which I will discuss below) can be distilled as intentional conduct causally linked to the crime, and in consideration of the mens rea issue in relation to JCE/aiding and abetting, I believe that a shift to a unitary model would be both possible and desirable. The superior who orders the pillaging of a village, the officer who plans it, the corporal who incites her troops to carry it out, the mayor of a nearby town who provide trucks to transport the loot, and the foot soldiers who remove valuables from civilian houses all contribute to the loss of property of the villagers. They are all to blame, and their responsibility can be assessed by examining whether they contributed to this loss with the required mens rea. The different degrees of responsibility of all these participants in the crime can be taken into account in sentencing, and, as Professor Stewart has argued, a summary of each convicted person’s criminal conduct could be included in the judgment’s disposition. On a more practical side, this system would also save considerable time in the drafting of international criminal judgments, where addressing multiple modes of liability requires considerable time and human resources. Since the costs of ICL are a constant concern across the international courts, together with the length of some proceedings, this practical advantage should be given just consideration.

A unitary model would not, in my view, affect an accused’s right to know the nature and cause of the charges against her, and consequently her ability to defend herself. While it is certainly indispensable – in fact, it is a fundamental right – for the accused to be informed of how she is alleged to have participated in a crime, international indictments often use rather general formulas such as “by way of her position, the accused ordered, instigated, planned, or otherwise aided and abetted…”. This practice, which very often includes charging suspects with all the available modes of liability, and then adjusting their aim during trial depending on the evidence, offers little information to the accused with respect to the actual nature of her causal contribution to the alleged crimes. On the other hand, it requires a diligent defence counsel to prepare for and address all the alleged modes. Moving away from modes of liability could require prosecutors to draft indictments which describe more precisely the accused’s alleged criminal conduct. It would also put an end to the practice of charging all modes of liability statutorily available, which in turn would allow the defence and the chambers to better focus their work and possibly shorten the duration of the proceedings.

Regarding superior responsibility, I see merits in the suggested separation of the failure to prevent from the failure to punish, treating the former as a form of participation and the latter as a separate crime. Failure to prevent crimes of subordinates which contributed, at least at the level of facilitating, the commission of the subordinates’ crimes may, if the accused acted with the required mens rea, amount to aiding and abetting by omission. In a unitary system, therefore, failure to prevent could be subsumed under a monistic model of participation which makes anyone who contributed to that crime with knowledge or intent responsible for it. As for failure to punish, the creation of a specific offence based on the commander’s dereliction of duty seems a preferable solution. First, it would avoid the theoretical difficulties encountered by judges and scholars when trying to introduce a causal nexus in this mode of liability. Second, on the front of general prevention, superiors would still be incentivized to properly exercise control and call to task criminal subordinates to avoid criminal sanctions, as well as administrative and disciplinary ones which may be imposed as accessory penalties for this type of offence. Finally, considering failure to punish as a discrete crime would be more respectful of general principles of culpability, which are violated when we convict a superior who negligently failed to punish the murder committed by his subordinate for murder (this violation is even more egregious if the superior did not have actual knowledge of the crime, but only “had reason to know” or, even worse, “should have known”).

Eliminating modes of liability, however, would certainly not be a panacea for blame attribution in ICL. In Italy, where there is a unitary system of participation, there have been long, vibrant, and often unresolved debates on different theories of causation and on the system’s adherence to the principle of culpability enshrined in the Constitution. Thus, the adoption of a unitary system in ICL would also presumably shift the debate and practical problems from modes of liability to the main pillars of blame attribution: causation and culpability. This shift would, I suspect, also be a healthy one for international criminal justice.

3.  Culpability distortions in the Italian system

With respect to culpability, Professor Stewart is perhaps too benevolent about the Italian unitary system’s ability to “avoid modes of liabilities acting as a prism that distorts responsibility, instead of assigning it in line with the culpability announced in the crime”.[8]

In Italy, the rules governing participation of a plurality of persons in the commission of a crime are set forth in articles 110 to 119 of the Penal Code. Article 110 essentially states that when a plurality of persons contribute to/participate in the commission of a crime, each of them is responsible for the sanction prescribed for that crime. The judge may then impose different sentences based on the intensity of their contribution and mens rea. To incur liability through article 110, a person must have at least “facilitated” the commission of the crime. As for the subjective element, article 110 requires either dolus or dolus eventualis (the appropriate definition of the latter is probably one of the most debated issues in Italian criminal law but, it suffices to say, dolus eventualis is generally interpreted as a higher standard than the foreseeability one in JCE III).

Article 116 of the Penal Code, however, states that when more people participate in the commission of a crime (for instance a bank robbery), if a different crime is committed by some of the participants (for instance murder), and this crime was not intended by the others, even those who did not intend the different crime will be punished for it, provided there exists a causal nexus between their conduct and the crime. If the crime not intended by some of the participants is graver than that agreed by them, the judge must take it into account at the sentencing stage. In spite of this mitigating factor, this is essentially a form of strict liability which violates the principle of culpability enshrined in the Italian Constitution.[9] The Constitutional Court partly corrected this state of affairs in 1965, stating that for a participant to incur liability for a crime different from the one she intended, the different crime must have been one that was at least a logically foreseeable development of the intended actions, according to common diligence and considering all the circumstances of the case. Thus, article 116 essentially allows a judge to convict someone for a crime requiring intent even when the evidence only establishes a much less culpable state of mind. This is one of the criticisms put forward by many scholars, including Professor Stewart, to JCE III. Those who defend article 116 might do so for reasons similar to those expressed by Professor Damaška in The Shadow Side of Command Responsibility (2001), where he stated that the persons to whom the graver crime is attributed through such system had already consciously decided to participate in a dangerous criminal activity, thus creating a “less dramatic shift in the register of culpability.”[10]

In sum, while I believe that a shift to a unitary model would be possible and could potentially have practical advantages for ICL, I am also aware that in the delicate field of criminal law there is no perfect system of blame attribution. Individual criminal responsibility is a fascinating but complex subject, and this comment of mine, based on years of practice may seem cursory from a scholarly perspective. I am aware that I have, in a non-scientific fashion, taken a position on issues which are the subject of vibrant academic debates, of long hours of legal research, and of thorough academic articles. Mine is the opinion of an Italian practitioner of international criminal justice, who in the past ten years has had the opportunity to test the advantages and limits of the pluralistic model of participation. I thank Professor Stewart for inviting me to share this practical perspective as part of this symposium.

 

 

 

 

[1] The views expressed in this comment are exclusively mine, and do not necessarily reflect the views of my present or former employers.

[2] While the jurisprudence of the ad hoc tribunals does not require a causal nexus between a commander failure to prevent and/or punish the crimes and their commission, I am aware that this position is criticized by scholars as non-reflective of customary international law. At the ICC, proof of a causal link is required, in terms of an increased risk, in relation to failure to prevent only, although some scholars and judges have expressed the view that causation is also required with regard to failure to punish.

[3] While ICTY jurisprudence cited by Professor Stewart in The End of Modes of Liability (p. 48) stated explicitly that proof of causal-effect relationship between the conduct of the aider and abettor and the commission of the crime is not necessary, it is possible that this was not meant to signify that no causal link at all is required between the acts of the aider and the crime. First, as noted by Stewart, the conduct of the aider or abettor needs to have a substantial effect on the commission of the crime. As a practitioner, I find it difficult to distinguish this requirement from causation. I am of the view that the Blaškić Chamber actually meant that the contribution of the aider or abettor need not be a conditio sine qua non. In the Applicable Law Section of the judgment in the case of Popović et al., in fact, paragraph 1018 first prescribes that the aider and abettor’s conduct must have had a substantial effect on the crime, and then clarifies that it is not necessary, however, to prove that the crime would not have been committed absent contribution of the aider and abettor.

[4] While significant is a lesser standard of contribution than substantive, I doubt that a review of the jurisprudence of the ad hoc tribunals would reveal a consistent and coherent practice in characterising the facts emerged from the evidence as a significant or substantive contribution.

[5] The End of Modes of Liability, pp. 38-39.

[6] To incur in JCE liability, a person must, inter alia, intend to further the common criminal plan.

[7] This problem is also noted by Matteo Costi in Introduzione al Diritto Penale Internazionale (terza edizione), Milano, Giuffrè, p. 117.

[8] The Strangely Familiar History of the Unitary Theory of Perpetration, p. 22.

[9] The Italian Constitution was issued in 1948, so when the Codice Rocco was issued in 1930 there was no constitutional impediment to article 116.

[10] In his article, Professor Damaška took as an example the felony murder rule in American law. It seems to me that the situation I described in the example features substantive similarity with this type of common law felony.

Norway: Three Codes, Three (Somewhat) Different Solutions


Jørn Jacobsen is a Professor in the Faculty of Law at the University of Bergen.


James G. Stewart’s argument for a unitary theory to replace the modes of liability in international criminal law is of particular interest from the point of view of Norwegian criminal law theory. Here, one of the most prominent contributors to the discipline, Bernhard Getz, made a similar claim in the latter half of the 19th century. Getz’ famous work on a unitary theory from 1875 is often hailed as a masterpiece (published in 1876 as Om den saakaldte Delagtighed i Forbrydelser – en strafferetlig Undersøkelse: Prøveforelæsning over selvvalgt Thema ved Concurrence om en Professorpost i Lovkyndighed). For good reasons too: he wrote it at only 25 years of age, and it certainly demonstrated a theoretical maturity that was a great surprise to Norway ‘s then fledgling criminal law discipline. After all, one counts Schweigaard’s commentaries from the 1840’s as the starting point for this discipline, and Getz’s work was the first significant theoretical contribution to it.

Getz became a professor the year after he published his book, and went on to have a huge impact on the formation of modern Norwegian criminal law. Getz and his close companion, Francis Hagerup, exhibited their impressive ambitions for criminal law in other manners too. They had international ambitions, and were both active in the AIDP (Association International de Droit Penal – International Association of Penal Law). Even more importantly from a Norwegian point of view, they filled key positions within the Norwegian political and legal order. Professor Hagerup served two separate terms as Prime Minister. Getz, for his part, held key positions in forming Norwegian legislation on criminal law and criminal procedure. He led both the commission preparing the Criminal Procedure Code of 1887 and the Criminal Code of 1902. He then became the first Director of Public Prosecutions in order to implement the Criminal Procedure Code of 1887. Certainly, this was the golden age for Norwegian criminal law – likely the first and only time when a criminal law professor held such prestigious positions in Norwegian society and with regard to the Norwegian criminal code, which was hailed as landmark legislation by central Continental criminal law scholars.

What then about Getz’s unitary theory? In itself, it was a critique of the then existing criminal code, the Criminal code of 1842. Norway’s first criminal code after independence was achieved in 1814. The code was itself a result of constitutionalization, as the Constitution of 1814 sect 94 required a criminal code to be enacted. This criminal code was imprinted by the Continental ideals at that time. The models used were the Code Penal (1810) and in particular the Code of Hannover (1840), which was itself inspired by Feuerbach’s Bavarian criminal code of 1813. Not surprisingly, this code differentiated between contributors to crime. In the code, a separate chapter was dedicated to ‘Participation’ (chp. 5), which was understood as something different from the ‘Perpetrator’. Here, the code had separate provisions for several different forms of participation, such as instigation of crime.

Starting from a concept of causation, Getz heavily criticized the 1842 Code. As there were no conceptual differences between the participator and the perpetrator, there was no reason to differentiate between them – thus the title ‘On so-called Participation in Crime’. The unitary theory was also the starting point for Getz when he embarked on the task of drafting the new Penal Code of 1902, which was celebrated throughout Europe. Here, in line with Getz’s program, there was no separate chapter on participation. In regard to sentencing, however, a section of the code in keeping with Stewart’s approach, assigned differences between different contributors’ importance. Getz clearly aimed at putting his theoretical enterprise into practice. However, the code of 1902 still included complicity as additional elements in a number of specific offenses. The additions do not appear to be consistently included in the code. This also left it for court practice to decide in regard to a number of offences whether complicity gave rise to criminal responsibility. In regard to several offences, the Supreme Court concluded that it did. Complicity therefore remained a central concept in Norwegian criminal law theory and practice even after Getz’s own code.

The lack of complete coherence between Getz theoretical project and his solution as a drafter of the code has often been emphasized in later literature. Unfortunately, Getz passed away at young age in 1901 – a year before his criminal code was enacted. As a consequence, we do not know how he would have responded to this criticism. Moreover, it is a point of intrigue for the current debate that later Norwegian scholars have not followed Getz’s approach on these questions. In particular, G. Astrup Hoel (1941) and Erling Johannes Husabø (1999) have criticized it. Husabø’s critique of Getz, in the most recent and extensive investigation into complicity in Norway, starts out from a different concept of causation and also offers conceptual arguments concerning the relation between ‘Perpetrator’ and ‘Participator’.

In 2015, the Norwegian code of 1902 was replaced by the Penal Code of 2005. At least in part, this legislative shift can be described as return to the original code of 1842. In the code of 2005, there is a general section on participation (sect. 15), which makes participation in crime in general subject to criminal responsibility unless otherwise stated in the offense. This new code does thereby not adopt the opposite solution to Getz’s code of 1902. In the preparatory works, there is no discussion of a unitary theory. Instead, treating complicity as a separate subject seems now to be taken for granted.

Lessons learned? There are two ways to see the rise and fall of the unitary theory in Norwegian criminal law. One way to see it is as an unfulfilled promise, one that was hindered by Getz’s early death, legal culture and other obstacles. The other way to see the Norwegian story is that the theory was flawed in the first place. There is not room for a detailed discussion of the subject here, nor on the particularities of international criminal law and the distinct challenges to modes of liability that this field of law faces. It seems clear, however, that the history of Norwegian criminal law at least is not a particularly strong argument for the potential of a unitary theory internationally.